Cross-language transfer is a core issue in second language acquisition field (Klein et al., 1986; Gass and Selinker, 1992). Behaviorism, cognition, and social schools differ in both their theoretical claims and empirical research findings regarding whether and how L1 influences second language acquisition (Cook and Singleton, 2014).

Behaviorism asserts that L1 is the primary cause of learners’ language acquisition difficulties and errors, with a greater difference between L1 and the target language leading to more difficulty learning the target language (Lado, 1957; Stockwell et al., 1965). However, the behaviorist viewpoint raises theoretical and empirical questions. From a theoretical perspective, many researchers argue that behaviorist theories of cross-language transfer ignore learners’ subjectivity, while viewing language acquisition as a stimulus-reflection habit-forming process (Krashen, 1985; Swain and Lapkin, 1995; Long, 1996). From an empirical perspective, numerous research findings demonstrate that cross-language differences do not necessarily lead to second language acquisition difficulties (Grauberg, 1971; Dulay and Burt, 1973), or serve as a main reason for second language acquisition difficulties (Kleinmann, 1978; Zobl, 1983; Pica, 1994).

The cognitive school’s two main branches, universal grammar theory and connectionist framework theory, emphasize the learner’s role as a cognitive subject in language acquisition (Cook and Singleton, 2014). The universal grammar theory posits two distinct views on whether L1 affects second language acquisition. One holds that universal grammar covers the second language’s initial state, and that L1 has no effect on second language acquisition (Bley-Vroman, 1990; Epstein, 1996; Platzack, 1996), while the other holds that L1 differs from the initial state of the second language, and that L1 affects second language acquisition (Schwartz and Sprouse, 1996; Vainikka and Young-Scholten, 1996). Both positions are supported by empirical studies. Different from the universal grammar theory, which emphasizes innate determinism, the connectionist framework theory holds that second language acquisition difficulty and learning speed largely depend on target language input frequency (Goldschneider and DeKeyser, 2001; Ellis, 2006), where L1 plays a role in regulating learners’ absorption of second language input. If the second language’s input information is similar to that of L1, a positive cross-language transfer occurs; if the second language’s input information differs from that of L1, a negative cross-language transfer occurs because the second language’s output is not what is expected. More importantly, at the elementary proficiency level of second language acquisition, learners usually draw on their L1 knowledge to compensate because their knowledge of the second language is not sufficient to allow them to fully express themselves. Therefore, learners at an elementary proficiency level are more likely to be affected by their L1, which also explains why second language acquisition development shows asymptotic and dynamic characteristics (Cook and Singleton, 2014).

Unlike the cognitive school, which views language as a psychological phenomenon, the sociocultural school focuses on both the influence of learners’ elements on second language acquisition and on how social and cultural aspects affect second language acquisition (Vygotsky, 1978; Ellis, 2008; Green and Abutalebi, 2013; Tong and Yip, 2015). Vygotsky’s sociocultural theory is influential in this school (Lantolf and Appel, 1994; Lantolf and Thorne, 2006), and posits that learners’ L1 serves as a mediation tool that helps them achieve their communicative purposes when learning a second language. Numerous cross-language transfer studies based on sociocultural theory also show that L1’s influence on learners’ second language acquisition varies with their second language proficiency; when learners are not proficient enough to control cognitive activity in the second language, they rely more on their L1 to complete challenging mental tasks (Frawley and Lantolf, 1985; Swain and Lapkin, 1998; Centeno-Cortés and Jiménez Jiménez, 2004).

In summary, various schools of cross-language transfer theory agree that L1 plays a role in second language acquisition (MacWhinney and Kroll, 2005). Most scholars acknowledge that cross-language transfer is a complex and dynamic cognitive process that is influenced by a variety of factors, including linguistics, psychology, and society, rather than a mechanical habit-forming process based on stimulus–response (Xu Q. et al., 2013). Therefore, L1’s role in second language acquisition needs to be examined from a systematic and developmental perspective (Larsen-Freeman, 1997; Larsen-Freeman and Cameron, 2008).

Chinese characters, vocabulary, and grammar are important components of Chinese language knowledge in relation to CSL acquisition. Chinese is regarded as a difficult language to learn, which is related to both the peculiarities of the Chinese writing system, and the peculiarities of Chinese vocabulary and grammar.

Chinese characters are the written symbols of Chinese and the basis for Chinese reading and writing. CSL learners, especially those from the non-Sinosphere, often find Chinese characters difficult to learn because of fundamental differences between Chinese characters and alphabets regarding stereoscopic structure and ideographic nature (Everson, 1998; Shi and Wan, 1998). First, compared with alphabets’ linear structure, Chinese characters have a square-shaped and more complex structure. Chinese character units comprise strokes and radicals, where strokes are the characters’ smallest units, and radicals are the characters’ secondary units, composed of strokes (Fei, 1996). Although there are few basic strokes in Chinese characters, the strokes have different deformations when in different positions in Chinese characters. For example, the stroke “” is “” in the Chinese character “快” (kùai, quick) and is “” in the Chinese character “水” (shuiˇ, water). In combining radicals into Chinese characters, the size and direction of strokes in each position also change with the radical’s position and the positional relationship between radicals. For example, the relative position and size of radical “口” in Chinese characters “扣”(kòu, button) and “器”(qì, vessel) differ. Second, Chinese characters belong to the morpheme-syllabic system, with a strong connection between form and meaning, but a poor connection with pronunciation (Tan et al., 2005). Chinese characters comprise four types: pictographic, ideographic, self-explanatory, and pictophonetic, which are all ideographic in nature. For example, the meaning of the pictographic character “刀” (daō, knife) is derived from the graphic “,” meaning knife; the meaning of the ideographic character “武”(wuˇ, military) comprises the meaning of two Chinese characters “止”(zhiˇ, stop) and “戈”(gē, dagger-ax), which means the cessation of war; the meaning of the self-explanatory character “刃”(rèn, blade) comprises the meaning of the pictographic character “刀” plus the indicator “,” which means knife blade; the pictophonetic character “湖”(hú, lake) has the semantic radical “,” indicating that the meaning of the character is related to water. Among the four types of Chinese characters, the pictophonetic character is the only one with a phonetic representation function (Li et al., 1992). However, the phonetic radical of the pictophonetic character can no longer accurately represent the pronunciation. For example, “触”(chù, touch) and “浊”(zhuó, muddy), which share the phonetic radical “虫” (chóng, insect), are not pronounced similarly. Chinese characters’ peculiarities make their acquisition tremendously difficult for CSL learners, especially for learners whose L1 is an alphabet script.

Vocabulary is the carrier of meaning, the key to understanding, and the basis of expression. Vocabulary knowledge includes both breadth knowledge and depth knowledge (Meara and Jones, 1988; Zareva et al., 2005; Moinzadeh and Moslehpour, 2012), where breadth knowledge refers to vocabulary size (Qian, 1999), and depth knowledge includes word aggregation relationships (such as synonymous relationships and context relationships) and combinatorial relationships (collocation relationships) (Nation, 1990). From a word aggregation perspective, first, vocabulary learning difficulties are reflected in differences in word concepts’ cognition in different languages. For example, when the Chinese use the word “鱼” (yú, fish) as a metaphor for people, it usually means dishonest, but in Russian, it means silent, and in Czech, it means robust. Second, the degree of word concept refinement differs across languages. Mandarin has a large number of synonyms, many of which have very little difference in meaning. For example: “承继” (chéngjì) and “继承” (jìchéng), both mean subsequent possession, but the former focuses on forward continuation, and the latter on backward continuation. CSL learners often struggle to grasp such subtle differences between words (Zhang, 2019). From a word combination relationships perspective, collocation knowledge reflects syntactic, semantic, and usage frequency information in the mental lexicon (Xing, 2013). Therefore, acquiring collocation knowledge includes mastering both target language collocation grammar rules and collocation words’ semantic category (Shi et al., 2021). However, each language has unique collocation rules, and many collocations are also based on various ethnic groups’ social psychology language habits, which cannot be logically explained (Yamashita and Jiang, 2010). For example, “black tea” in English corresponds to “红茶”(hóngchá) rather than “黑茶”(hēichá) in Chinese. This difference in collocations across languages poses another difficulty in CSL learners’ vocabulary acquisition.

Grammar is the organizational rule of a language; an important sign of second language mastery is mastering its grammar (Li, 2016). Chinese grammar is difficult to learn because, first, unlike Indo-European languages that use a rich variety of morphology to express grammatical relationships, Chinese is an isolated language and lacks strict morphology, so word order and function word are important means of expressing grammatical relationships (Lv, 1979). The grammatical meaning often differs according to the word order. For example, in “我看” (woˇkàn, I see) and “看我” (kànwoˇ, see me), the former indicates a subject-predicate relationship, while the latter indicates a verb-object relationship. The choice of whether to use function words, and the use of different function words, often indicates different grammatical relationships. For example, “买书”(mǎishū) is a verb-object relationship, while “买的书”(mǎideshū) is a subordinate relationship. Second, unlike in English and many other languages, Chinese has no one-to-one correspondence between word class and syntactic constituents (Zhu, 1985)—a word class can act as multiple syntactic constituents, and a syntactic constituent can also be acted on by multiple word classes. Chinese also has some syntactic peculiarities that set it apart from other subject-verb-object languages, such as prepositional phrases followed by verbs, and relative clauses placed before the head. Additionally, Chinese contains unique syntactic constructions, such as pivotal and ba-structure sentences. Therefore, Chinese grammar acquisition difficulties may vary for CSL learners with different L1 backgrounds, owing to these distinctive aspects of Chinese grammar.

However, although Chinese characters, vocabulary, and grammar have distinct characteristics, this does not mean that they are unrelated. From the perspective of Chinese language research, Chinese characters belong to the morpheme-syllabic script, meaning that Chinese characters are not only syllables that represent pronunciation, but also words or morphemes that represent meanings (Li, 2009). The integration of Chinese characters’ form, pronunciation, and meaning has exerted great influence on Chinese words’ syllable form, formation, and meaning composition. Additionally, as Chinese word construction is similar to phrase or sentence construction, syntactic structures and words share a selective restriction relationship (Shi and Yang, 2021); therefore, Chinese character, vocabulary, and grammar characteristics often influence each other. In addition, evidence from many empirical studies shows that knowledge of Chinese characters’ sublexical and grammatical features is also activated during Chinese character processing (Tsai et al., 2004; Yan et al., 2009, 2012; Tsang et al., 2017; Yeh et al., 2017; Pan et al., 2019), which indicates that CSL learners learn Chinese characters, vocabulary, and grammar simultaneously.

In sum, language systems interact across levels, and Chinese character, vocabulary, and grammar characteristics and their acquisition often influence each other. Only when CSL learners’ Chinese characters, vocabulary, and grammar are all well-developed can Chinese proficiency improve. Therefore, these three elements should not be separated when examining CSL learners’ language knowledge acquisition.

To examine whether and how linguistic distance affects Chinese character acquisition by elementary, intermediate, and advanced proficiency CSL learners, we analyzed Chinese character test scores from 2,250 CSL learners in nine groups. Table 2 shows the Chinese character test score descriptive statistics for the nine CSL learner groups.

Two-way between-subjects ANOVAs revealed a significant main effect of linguistic distance (F_(2，2,241) = 51.69, p < 0.001, η_p² = 0.044) and Chinese proficiency (F_(2，2,241) = 1102.23, p < 0.001, η_p² ₌0.496), with a significant interaction effect (F_(4，2,241) = 2.86, p = 0.02, η_p² = 0.005). To further explore specific differences between the groups, we conducted a simple effects test, which showed that at the elementary Chinese proficiency level, the short-distance group Chinese character scores were significantly higher than those of the middle-distance group (mean difference = 0.69, p = 0.02) and long-distance group (mean difference = 0.72, p = 0.01). Although the middle-distance group’s Chinese character scores were higher than those of the long-distance group, the difference was not significant (mean difference = 0.024, p = 1.00). At the intermediate Chinese proficiency level, the short-distance group’s Chinese character scores were significantly higher than those of the middle-distance group (mean difference = 1.36, p < 0.001); the short-distance group scores were also significantly higher than those of the long-distance group (mean difference = 1.80, p < 0.001). Although the middle-distance group’s performance was higher than that of the long-distance group, the difference was not significant (mean difference = 0.44, p = 0.23). At the advanced Chinese proficiency level, the short-distance group’s Chinese character scores were significantly higher than those in the middle-distance group (mean difference = 1.43, p < 0.001); the short-distance group’s Chinese character scores were also significantly higher than those in the long-distance group (mean difference = 1.64, p < 0.001), but there was no significant difference between the middle-distance group and the long-distance group (mean difference = 0.21, p = 0.79; see Figure 2).

The findings from Quasi-Experiment 1 showed how linguistic distance affected CSL learners’ Chinese character acquisition, where performance differed across the elementary, intermediate, and advanced Chinese proficiency conditions, but the short-distance group’s Chinese character performance was noticeably better than that of the middle-and long-distance groups. Therefore, differences in linguistic distance’s influence on CSL learners’ Chinese character acquisition predominately manifested between the short-and middle-distance groups, and between the short-and long-distance groups.

To examine whether and how linguistic distance affects Chinese vocabulary acquisition by elementary, intermediate, and advanced proficiency CSL learners, we analyzed HSK vocabulary test scores from 2,250 CSL learners in nine groups. Table 3 shows Chinese vocabulary test score descriptive statistics for the nine CSL learner groups.

Two-way between-subjects ANOVAs revealed a significant main effect of linguistic distance (F_(2，2,241) = 43.41, p < 0.001, η_p² = 0.037) and Chinese proficiency (F_(2，2,241) = 2135.19, p < 0.001, η_p² = 0.656), with a significant interaction effect (F_(4，2,241) = 2.86, p = 0.02, η_p² = 0.005).

A simple effect test showed that for elementary Chinese proficiency, the short-distance group’s vocabulary scores were significantly higher than those of the middle-distance group (mean difference = 1.13, p < 0.001), and those of the long-distance group (mean difference = 1.38, p < 0.001). However, there was no significant difference between the middle-distance and long-distance groups (mean difference = 0.25, p = 0.67). For intermediate Chinese proficiency, the short-distance group’s vocabulary scores were significantly higher than those of the middle-distance (mean difference = 1.05, p < 0.001), and long-distanc (mean difference = 1.62, p < 0.001) groups. However, there was no significant difference between the middle-distance and long-distance groups’ scores (mean difference = 0.58, p = 0.06). For advanced Chinese proficiency, there was no significant difference between the short-distance and middle-distance groups’ vocabulary scores (mean difference = 0.10, p = 0.97), but the two groups’ scores were significantly higher than those of the long-distance group (mean difference = 0.98, p < 0.001, and mean difference = 0.88, p = 0.001, respectively) (see Figure 4).

Quasi-Experiment 2 showed that linguistic distance affected CSL learners’ vocabulary acquisition, and that vocabulary performance differed across distance groups under the elementary, intermediate, and advanced Chinese proficiency conditions. First, the short-distance group’s vocabulary scores were significantly higher than those of the middle-distance group in the elementary and intermediate Chinese proficiency conditions, but there was no significant difference between the two groups in the advanced Chinese proficiency condition. This indicates that difference in vocabulary acquisition performance between the short-and middle-distance groups decreased as Chinese proficiency increased. Second, there was no significant difference between the middle-distance and long-distance groups in the elementary and intermediate proficiency conditions, but the middle-distance group had significantly higher vocabulary scores than the long-distance group under the advanced Chinese proficiency condition. This indicates that as Chinese proficiency improved, vocabulary acquisition performance differences between the middle-distance and the long-distance groups increased. Finally, although the difference in vocabulary scores between the short-and long-distance groups showed a decreasing trend, the short-distance group’s vocabulary scores were significantly higher than those of the long-distance group at any Chinese proficiency level.

To examine whether and how linguistic distance affected Chinese grammar acquisition by elementary, intermediate, and advanced proficiency, we analyzed the HSK grammar scores of 2,250 CSL learners in nine groups. Table 4 shows the grammar test score descriptive statistics for the nine groups.

Two-way between-subjects ANOVA revealed a significant main effect of linguistic distance (F_(2，2,241) = 16.111, p < 0.001, η_p² = 0.014) and Chinese proficiency (F_(2，2,241) = 2499.552, p < 0.001, η_p² = 0.690), and a significant interaction effect (F_(4，2,241) = 3.323, p = 0.01, η_p² = 0.006).

The simple effect test showed that the elementary proficiency group had no significant difference in grammar scores across the long-, middle-, and short-distance groups (mean difference = 0.58, p = 0.23; mean difference = 0.43, p = 0.49; mean difference = 0.15, p = 0.96, respectively). The intermediate proficiency group’s long-distance grammar scores were significantly higher than those of the short-distance group (mean difference = 1.03, p = 0.01). The middle-distance group’s grammar scores were also significantly higher than those of the short-distance group (mean difference = 0.86, p = 0.03), but there was no significant difference between the long-and middle-distance groups (mean difference = 0.17, p = 0.94). For advanced proficiency, the long-distance group’s grammar scores were significantly higher than those of the short-distance group (mean difference = 1.23, p = 0.001); the middle-distance group’s grammar scores were also significantly higher than those of the short-distance group (mean difference = 1.83, p < 0.001). Although the middle-distance group’s average score was higher than that of the long-distance group, the difference was not significant (mean difference = 0.60, p = 0.20) (see Figure 6).

Quasi-Experiment 3 showed that linguistic distance affected CSL learners’ grammar acquisition, and that distance groups’ grammar scores were not consistent across the elementary, intermediate, and advanced proficiency groups. Under elementary Chinese proficiency, there was no significant difference in grammar scores across the distance groups. However, under intermediate and advanced proficiency, the middle-and long-distance groups’ grammar scores were significantly higher than those of the short-distance group, indicating that CSL grammar acquisition differences increased as Chinese proficiency improved, as shown between the long-and short-distance groups, and the middle-and short-distance groups.

Quasi-Experiment 1 examined linguistic distance effects on CSL learners’ Chinese character acquisition at different Chinese proficiency levels. The results showed that, for elementary, intermediate, or advanced Chinese proficiency levels, Chinese character acquisition was more favorable with a shorter linguistic distance between L1 and Chinese, but this effect was only observed between the short-and middle-distance groups and the short-and long-distance groups. There were six L1s in the short-distance group comprising Vietnamese, Thai, Korean, Indonesian, Yoruba, and Japanese individuals. Japan, Korea, and Vietnam are Sinosphere member countries, while Indonesia and Thailand are both neighbors of China. Thus, our findings are consistent with the findings of most previous studies that Chinese characters are better acquired in Sinosphere CSL learners than in non-Sinosphere CSL learners (Jiang, 2003; Feng et al., 2005; Zhang, 2017; Loh et al., 2018; Deng and Hu, 2022). This result is partially consistent with cross-language transfer theory assumptions. According to previous studies, CSL learners in short-distance groups (mainly from the Sinosphere) have a basic understanding of the rules for writing Chinese characters because their native script has Chinese character forms, and they can acquire Chinese character meanings much easier because they have more frequent exposure to Chinese culture. Therefore, short-distance CSL learners have an advantage in both Chinese character experience and literacy (Feng et al., 2005; Feng, 2006; Lin and Collins, 2012). By contrast, the middle-and long-distance CSL learners have alphabetic language backgrounds; the stereoscopic structure of Chinese characters is more complex compared to the linear structure of alphabetic characters (Everson, 1998; Shi and Wan, 1998). Additionally, literary strategies in relation to Chinese characters differ greatly from those of alphabetic characters (Tan et al., 2005). The form-phonemic connection of Chinese characters is poor; however, CSL learners from alphabetic backgrounds are often influenced by their L1 and tend to use native-like form-phonemic strategies to recognize Chinese characters (Jiang, 2003). Moreover, there is a long-standing lack of emphasis on Chinese characters when teaching CSL (Li, 2017). All these factors increase the difficulty of Chinese character acquisition for CSL learners from non-Sinosphere countries, resulting in their inability to avoid the influence of negative cross-language transfer on Chinese character acquisition, even if they reach a relatively advanced Chinese proficiency level.

Quasi-Experiment 2 examined the influence of linguistic distance on CSL learners’ vocabulary acquisition at different proficiency levels. The results showed that CSL learners with elementary and intermediate proficiency and short linguistic distance had a vocabulary acquisition advantage, and that CSL learners with advanced proficiency and either middle or short linguistic distance showed significantly better performance than those with a long linguistic difference. First, these results are partially consistent with the finding that shorter linguistic distance is more favorable to vocabulary acquisition, which accords with cross-language transfer theory and most previous study findings (Zhang, 2006; Cai, 2017; Luo and Duan, 2019; Chang, 2020). Previous studies have shown that, because Sinosphere L1s share many Chinese origin words and morphemes, CSL learners from Sinosphere countries have a certain awareness of Chinese morphemes and can use morphemic strategies to learn new words at the elementary stage (Hong, 2011; Xu, 2014). Shorter linguistic distance to Chinese often means more exposure to Chinese culture; since words in different languages can be linked through concepts (Zhang et al., 2011), shorter linguistic distance also means easier conceptual word linkage, which makes positive cross-language transfer easier.

Second, vocabulary scores of all distance groups significantly increased with improvements in CSL learners’ Chinese proficiency levels. Vocabulary score differences between the short-and middle-distance groups significantly decreased, but vocabulary score differences between the long-and middle-distance groups significantly increased. We believe this result reflects a difference in the linguistic distance effect on vocabulary acquisition speed for non-Sinosphere CSL learners. As the middle-distance group had a linguistic distance advantage over the long-distance group, vocabulary knowledge expansion with improved Chinese proficiency could reduce the effect of negative cross-language transfer; whereas CSL learners in the long-distance group, whose L1 differed more from Chinese, were more affected by negative cross-language transfer, requiring them to overcome more vocabulary acquisition difficulties, which resulted in significantly slower vocabulary acquisition speeds for CSL learners in the long-distance group compared to the middle-distance group. Combined with the trend (although not significant) for decreased vocabulary score differences between the short-and long-distance groups, we believe that linguistic distance’s influence on CSL learners’ vocabulary acquisition generally diminished with increased Chinese proficiency.

Additionally, in comparing the results with those of Quasi-Experiment 1, we found similarities in the effects of linguistic distance on Chinese character and vocabulary acquisition, where shorter linguistic distance between L1 and Chinese was associated with more favorable acquisition, and showed a significant acquisition advantage for Sinosphere CSL learners at any proficiency level. This is most likely a result of Chinese characters and words being similarly acquired because of their blurred boundaries (Li, 2009). However, there were also differences in linguistic distance effects on character and vocabulary acquisition. Quasi-Experiment 2 showed no significant difference in vocabulary performance between the short-and middle-distance groups for CSL learners with advanced Chinese proficiency, while Quasi-Experiment 1 showed a significant difference. Considering that the HSK Chinese character subtest examines learners’ Chinese character or morpheme writing, and the vocabulary subtest focuses on word or morpheme recognition, this result supports, to some extent, the finding that Chinese character or morpheme writing is more influenced by linguistic distance than is reading (Wu et al., 2006; Liu, 2008; Li et al., 2014). Based on previous studies, we believe that the main reason for this phenomenon is that Chinese character recognition can be accomplished by using only part of a character’s information to remember its pronunciation or meaning, so it is only necessary to break these characters down into strokes; whereas writing requires learners to recall and reproduce the characters through their pronunciation and meaning, which requires mastery of not only the character radicals, but also the smaller stroke units (Jiang, 2000; Yeh et al., 2003; Xu Y. et al., 2013; Ke and Chan, 2017). Therefore, writing Chinese characters or morphemes is more difficult, resulting in a more profound negative cross-language transfer effect on Chinese character writing.

Interestingly, Quasi-Experiment 3 showed a pattern of cross-language transfer effects that completely differed from those found in Chinese character and vocabulary acquisition. There was no significant linguistic distance effect on grammar acquisition for CSL learners with elementary Chinese proficiency. This may be because cross-language transfer is evident in early stages, but in the middle-and long-distance groups, linguistic differences’ promoting effect weakened the L1 negative transfer effects and facilitated acquisition (Yuan, 2004; Wu and Zhao, 2018). Meanwhile, numerous studies have demonstrated that elementary proficiency learners rely more on lexical semantic information than syntactic information (Clahsen and Felser, 2006, 2018; Rah and Adone, 2010); therefore, grammatical knowledge development frequently lags behind that of content meaning knowledge, such as Chinese characters and vocabulary, which may be another reason why L1 differences are unlikely to show a significant grammar acquisition effect for CSL learners with elementary Chinese proficiency.

Although linguistic distance influenced CSL learners’ grammar acquisition at intermediate and advanced proficiency levels, short distance did not contribute to grammar acquisition at these proficiency levels. In other words, short distance did not produce a positive cross-language transfer, but rather had a hindering effect.

Previous studies suggest that this result may stem from a greater difference between learners’ L1 and Chinese, which makes learners more likely to pay attention to the differences, facilitating acquisition (Yuan, 2004; Wu and Zhao, 2018); or it could be that a high degree of cross-language similarity makes learners more likely to ignore differences, making acquisition more difficult (Laufer, 1990; Ellis, 1999). This result can only occur when language features’ similarity interference effects are significantly stronger than language differences’ facilitation effects, and significantly stronger than L1 transfer effects. Additionally, with increased grammar learning content, the probability of overgeneralizing will be greater (Ellis, 1994); compared with obvious differences between learners’ L1 and target language, the probability of overgeneralizing subtle differences between learners’ L1 and target language is greater (Ellis, 1999). In other words, as Chinese proficiency improves and CSL learners are exposed to more grammatical knowledge, short-distance group CSL learners are not only subject to interference from certain grammar with highly similar features to Chinese, but also to interference from their native grammar. This may result in a longer period of confusion before learners fully master Chinese grammar. Therefore, it is reasonable to believe that enhanced interference effects of short-distance L1 on CSL learners’ grammar acquisition inhibits grammatical competence development as their Chinese proficiency improves.

Previous studies showed that avoidance strategies and psycho-typological distance weaken L1 negative transfer and result in a non-significant acquisition advantage for short-distance CSL learners. However, as the HSK is a high-risk test with objective items, examinees are less likely to use avoidance strategies. The role of learners’ perceived differences in language typology is beyond the scope of our study; these issues can be further explored in future studies.

Our study also found no significant differences in Chinese character, vocabulary, and grammar scores between different distance groups at each Chinese proficiency level, except for vocabulary scores for CSL learners with advanced Chinese proficiency. This finding indicates that linguistic distance is important for distinguishing language knowledge acquisition performance in both Sinosphere and non-Sinosphere CSL learners, but does not play an obvious role in distinguishing language knowledge acquisition performance within non-Sinosphere CSL learners. Given the lack of studies that specifically focus on differences in Chinese acquisition within non-Sinosphere CSL learners, we cannot yet offer a general explanation.

Our results show that linguistic distance has different patterns of influence on Chinese character, vocabulary, and grammar acquisition. As Ellis (1999) asserted, L1’s influence on second language acquisition is not always a one-way positive or negative process, but involves a reciprocally dynamic process (Xu Q. et al., 2013).

Previous studies showed that shorter linguistic distances between L1 and Chinese facilitate CSL learners’ language skill development, such as listening, speaking, reading, and writing (Wang and Cui, 2018; Wang and Yang, 2019; Xu and Zi, 2020). Using HSK test data, we investigated the impact of linguistic distance between L1 and Chinese on character, vocabulary, and grammar knowledge acquisition. The results showed differences in linguistic distance’s effect on the total characteristics of three types of language knowledge, and differences in the degree of influence on three types of language knowledge acquisition at different Chinese language proficiency levels. Our findings indicate that, at least under the influence of linguistic distance, language knowledge acquisition is not consistent with language skills acquisition, because the language knowledge acquisition process is more complex (Ding and Xiao, 2016; Wu, 2017).